Hacker News Reader

If you fed all of the Sherlock Holmes stories from archive.org into this, how many new stories at a comparable level would be spat out from it?

I’d be shocked if those stories weren’t already in GPT-4’s training data.